BFF Figure Creation

[SeanDuHare]

This has been a long-standing hope for BFF & Antoine has a specific use case we can build from,

This is a NON-URGENT message:
I was thinking of a functionality idea for BFF/Vol-E, and I don't know where to post it.
One thing we're doing more and more internally is panels of antibodies. For those, being able to compare different files side by side is even more important. I know this is a huge project, so I’m not saying we should do it, but I’d like to add to the list of future ideas: a “figure maker in BFF.”
In this tool, you could select data in BFF—for example, “Experiment X / Antibody Y”—and ask BFF to generate a grid view. And you could pick other metadata to organize the grid. For example asking that the first column would display all the data for ‘Timepoint 1’, the second column for ‘Timepoint 2’, and each row would represent a different ‘Gene’ from the dataset.
The most incredible part would be if we could still interact with each panel in that figure the same way we can interact with a single file in Vol-E 😊.
For us (and many other biologists), it would be a powerful way to explore biological data and gain insight quickly. It could also enable full reproduction of a figure from a paper—with the added bonus of interactivity on the underlying data 😊

Example of desired thing below,

[will-moore]

Can I suggest that a nice solution here would be to connect BFF and OMERO.figure?

Once OMERO.figure can support OME.zarr (work in progress at ome/omero-figure#619) then BFF could be taught to generate the OMERO.figure JSON file to display a figure similar to the one above.
The only "missing step" in that workflow is how to pass that JSON file to OMERO.figure, since BFF can't host it if it's generated on the fly.

Demo of current work from the PR above: https://will-moore.github.io/omero-figure/?file=https://gist.githubusercontent.com/will-moore/fe0e260544b46af6e1e523b288fc85bc/raw/30547e61d4d8753ef0016f0a70435f1aafb43c2f/OMERO.figure_NGFF_demo.json

[SeanDuHare]

@will-moore hey will, checking back in on this. Is OMERO.figure something anyone can use? As in, can anyone, including those without OMERO, use OMERO.figure? If so, is there some public OMERO.figure website I can point BFF at? If it is private - let me know if you think it would be useful for us to implement this for OMERO users.

[will-moore]

Hi - thanks for the ping. Yes, you can certainly use OMERO.figure without OMERO. We are actually doing a workshop on this today at OME 2026: https://downloads.openmicroscopy.org/presentations/2026/OME-Duesseldorf/
One of the discussion points is whether we should rename OMERO.figure (e.g. to OME figure) to avoid this confusion in future?

Have a look at:
https://ome.github.io/omero-figure/?file=https://gist.githubusercontent.com/will-moore/fe0e260544b46af6e1e523b288fc85bc/raw/30547e61d4d8753ef0016f0a70435f1aafb43c2f/OMERO.figure_NGFF_demo.json

That's the static app, deployed from the master branch, loading a figure from this gist.

I seem to remember hearing that you can open a bunch of images in VolE by pushing the state into localStorage where VolE can read from? That sounds like a great solution if we could get that to work with figure too?

[SeanDuHare]

That is fantastic! IMO the name does sound like it is a plug-in to OMERO rather than a standalone app.

I'm walking through the workshop demo and this is very handy! I have a better idea of how OMERO.figure is intended to work now I think.

I seem to remember hearing that you can open a bunch of images in VolE by pushing the state into localStorage where VolE can read from? That sounds like a great solution if we could get that to work with figure too?

@toloudis could explain this better but yes this is something we try to do a for a couple of apps & I could see being useful for OMERO.figure as well :)

[frasercl]

I seem to remember hearing that you can open a bunch of images in VolE by pushing the state into localStorage where VolE can read from? That sounds like a great solution if we could get that to work with figure too?

This document describes the protocol from Vol-E's perspective: EXTERNAL_APP_MESSAGE_SPEC.md.

The current iteration actually no longer uses localStorage as part of the contract between BFF and Vol-E (we were ultimately foiled by browser security policies), though Vol-E stores the image collection it receives so that users can return to it later. The process now basically looks like:

1. BFF opens a new Vol-E window with a special query parameter that says "I've got more data to send to you"
2. Vol-E postMessages back to BFF as soon as it loads to say "I'm ready to receive more data"
3. BFF postMessages the data back.

Relevant PRs are #621 here and allen-cell-animated/vole-app#467 on the Vol-E side.

Now that the functionality is there in BFF, it shouldn't be too hard to get it talking to OMERO.figure as well, as long as figure is willing to implement a similar contract to the one described in the document above. (Just the steps — the content of the message can be whatever figure wants!)

[SeanDuHare]

@will-moore Hey Will, checking back in. How would you envision OMERO.figure integrating with BFF? It seems the information for the figure is stored in a JSON. Would BFF be able to post a request to create a blank JSON (or fill it in with Zarrs) and the user go from there?

[will-moore]

Hi, I can imagine a couple of ways to integrate.

• Simple Open-with: Just select 1 or more images in BFF and Open-with OMERO.figure -> e.g. https://ome.github.io/omero-figure/?image=https://livingobjects.ebi.ac.uk/idr/zarr/v0.5/idr0062A/6001240_labels.zarr&image=https://livingobjects.ebi.ac.uk/idr/zarr/v0.4/idr0048A/9846151.zarr/0
• Full figure creation in BFF: The figure json format is described at https://github.com/ome/omero-figure/blob/master/docs/figure_file_format.rst#json-format - although that needs updating with support for OME-Zarr images. You can see an example of a figure with OME-Zarr images (the figure above) at https://gist.githubusercontent.com/will-moore/fe0e260544b46af6e1e523b288fc85bc/raw/30547e61d4d8753ef0016f0a70435f1aafb43c2f/OMERO.figure_NGFF_demo.json - I can't guarantee that won't change, since OME-Zarr support is quite new in figure, but will try to keep it to a minimum/non-breaking if possible. We have discussed using OME-Zarr collections (RFC-8) as a new format for figure.json but that still seems a long way off so we'll not wait for that.

I'm kinda wondering how the UI & workflow might look in BFF. In the example figure above, you've got a grid of images with 2 parameters: Time-points (x 9) and treatments (x 8). You could imagine in BFF that you have 72 separate images (or use a Filter to give you 72 images out of a bigger collection), then Group By > Timepoint and Group By > Treatment - so you have 1 image in each group. Then you could possibly use that info to generate a figure json that represents the figure above.
However, it's probably quite tricky to give users a good idea of how that works and what to expect without creating a Preview of some sort? You could easily end up with multiple images in some Groups and none in others. I guess that might not matter too much - you could sort out various issues in Figure app itself.

• We might be able to get away with something in-between those two, whereby we pass the info we need in the request parameters (as in the first case) but we add a little bit more essential info such as labels to the request parameters. But this could quickly get very complex and exceed the typical request size limit.

I'm certainly happy to implement the postMessage handshake as and when we get to that point. In the meantime, if you generate figure.json in BFF, you can open it in figure at https://ome.github.io/omero-figure/ by doing File > Import from JSON... and paste it in.